mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-31 19:16:29 +00:00
39975613b1
Second, deeper pass over tools/gateway/hermes_cli plus first pass over the trees wave 1 missed (acp, acp_adapter, skills, computer_use, docker, dashboard, conformance, monitoring, secret_sources, hermes_state, providers). Same rubric as wave 1 (AGENTS.md test policy); security, alternation/caching invariants, issue-number regressions, and E2E kept. Real test-quality fixes found and rooted out along the way: - tests/tools/test_command_guards.py made real auxiliary-LLM HTTPS calls (DEFAULT_CONFIG smart-approval leaked in) — pinned approval mode=manual via autouse fixture: 17.4s → 0.4s. - test_model_switch_custom_providers.py / test_user_providers_model_switch.py silently probed live provider catalogs (~2s/test) — stubbed cached_provider_model_ids/provider_model_ids/fetch_api_models. - test_telegram_noise_filter.py: 15-platform copy-paste matrix over shared gateway.run logic → 3 representative platforms (55s → 3.9s). - test_gateway_shutdown.py: stop()'s 5s interrupt-deadline loop spun on MagicMock agents — interrupt.side_effect now clears _running_agents (22s → 1.0s). - test_gateway_inactivity_timeout.py poll-harness timings shrunk 3-5x (24s → 1.1s); test_mcp_stability.py backoff/SIGTERM-grace sleeps patched (15.4s → 2.5s); test_async_delegation.py negative-drain wait 5s → 0.5s. - test_telegram_init_deadline.py: loop-block margin restored to 1.0s with rationale comment — the watchdog-dump assertion needs the loop blocked well past deadline+grace under parallel load (flaked once in the 40-worker verification run at a 0.2s margin). Verification: full hermetic suite via scripts/run_tests.sh — 2,438 files, 21,718 tests passed, 0 failed, 293.9s wall. Suite totals vs original baseline: 46,820 → 19,757 test functions (−57.8%), wall 583.5s → 293.9s (−50%), subprocess CPU 13,564s → 11,623s.
78 lines
3.2 KiB
Python
78 lines
3.2 KiB
Python
"""Tests for subagent summary budgeting (PR #9126).
|
|
|
|
delegate_task caps subagent summaries against the parent's remaining context
|
|
headroom (split across the batch) before they enter the parent's context, and
|
|
spills the full text to disk so nothing is lost. This guards the
|
|
compression/429 death spiral that batch fan-out could trigger by returning N
|
|
full summaries verbatim into the parent.
|
|
"""
|
|
|
|
import os
|
|
import tempfile
|
|
|
|
import pytest
|
|
|
|
import tools.delegate_tool as dt
|
|
|
|
|
|
class _FakeCompressor:
|
|
def __init__(self, context_length, max_tokens):
|
|
self.context_length = context_length
|
|
self.max_tokens = max_tokens
|
|
|
|
|
|
class _FakeParent:
|
|
def __init__(self, context_length, used_tokens, max_tokens):
|
|
self.context_compressor = _FakeCompressor(context_length, max_tokens)
|
|
self.session_prompt_tokens = used_tokens
|
|
|
|
|
|
def test_small_summaries_pass_through_untouched():
|
|
parent = _FakeParent(context_length=200_000, used_tokens=10_000, max_tokens=8_000)
|
|
results = [
|
|
{"task_index": 0, "summary": "short result A", "status": "completed"},
|
|
{"task_index": 1, "summary": "short result B", "status": "completed"},
|
|
]
|
|
dt._apply_summary_budget(results, parent)
|
|
assert results[0]["summary"] == "short result A"
|
|
assert "summary_truncated" not in results[0]
|
|
assert "summary_truncated" not in results[1]
|
|
|
|
|
|
def test_batch_overflow_trimmed_and_spilled_losslessly(monkeypatch):
|
|
# Isolate spill directory to a temp HERMES_HOME.
|
|
with tempfile.TemporaryDirectory() as td:
|
|
monkeypatch.setenv("HERMES_HOME", os.path.join(td, ".hermes"))
|
|
# Distinct head + tail markers so we can prove the tail survives.
|
|
big = "HEAD_MARKER\n" + ("X" * 50_000) + "\nTAIL_MARKER"
|
|
# Parent nearly full (120k/131k) → tiny headroom → aggressive trim.
|
|
parent = _FakeParent(context_length=131_000, used_tokens=120_000, max_tokens=8_000)
|
|
results = [
|
|
{"task_index": i, "summary": big, "status": "completed"} for i in range(5)
|
|
]
|
|
dt._apply_summary_budget(results, parent)
|
|
for r in results:
|
|
assert r["summary_truncated"] is True
|
|
assert len(r["summary"]) < len(big)
|
|
# Head+tail window: both ends survive in-context.
|
|
assert "HEAD_MARKER" in r["summary"]
|
|
assert "TAIL_MARKER" in r["summary"]
|
|
path = r.get("summary_full_path")
|
|
assert path and os.path.exists(path)
|
|
# The spill file holds the FULL original text — nothing is lost.
|
|
with open(path, encoding="utf-8") as fh:
|
|
assert fh.read() == big
|
|
# The footer points the parent at the full version with an offset.
|
|
assert "read_file" in r["summary"]
|
|
assert "offset=" in r["summary"]
|
|
# Spilled into the delegation cache (mounted into remote backends).
|
|
assert os.path.join("cache", "delegation") in path
|
|
|
|
|
|
def test_empty_results_is_noop():
|
|
# No summaries → nothing to do, must not raise.
|
|
dt._apply_summary_budget([], _FakeParent(131_000, 1_000, 8_000))
|
|
dt._apply_summary_budget(
|
|
[{"task_index": 0, "status": "failed", "summary": None}],
|
|
_FakeParent(131_000, 1_000, 8_000),
|
|
)
|